Design of a Total Pressure Distortion Generator for Aircraft Engine Testing by Kevin B. Cramer Thesis submitted to the Faculty of the Virginia Polytechnic Institute and State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE in Mechanical Engineering APPROVED: _____________________________ _____________________________ W.F. O’Brien, Committee Co-Chair P.S. King, Committee Co-Chair _____________________________ C.L. Dancey, Committee Member May 2002 Blacksburg, Virginia Keywords: distortion, HCF, surge margin, stall, generator, jet engine, non-uniform
112
Embed
Design of a Total Pressure Distortion Generator for ...characteristics, a mechanism must be developed that when actuated, can change the distortion pattern and intensity with respect
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Design of a Total Pressure Distortion Generator for Aircraft Engine Testing
by
Kevin B. Cramer
Thesis submitted to the Faculty of the Virginia Polytechnic Institute and State University
in partial fulfillment of the requirements for the degree of
Design of Total Pressure Distortion Generator for Aircraft Engine Testing
by
Kevin B. Cramer
Committee Co-Chair: W.F. O’Brien
Committee Co-Chair: P.S. King
Mechanical Engineering
(ABSTRACT)
A new method and mechanism for generating non-uniform, or distorted, aircraft
engine inlet flow is being developed in order to account for dynamic changes during the
creation and propagation of the distortion. Total pressure distortions occur in gas turbine
engines when the incoming flow is disturbed. Dynamic total pressure changes may
happen slowly, or may occur very rapidly. The disturbance of the incoming flow can
change engine operating characteristics, including lowering the surge limit and creating
High Cycle Fatigue incidents. In order to create a distorted flow with dynamic
characteristics, a mechanism must be developed that when actuated, can change the
distortion pattern and intensity with respect to time.
This work covers the initial design of both the distorting and actuating device.
The design chosen is a low force design that is practically independent of flow forces.
This allows the system to be easily sized for all flow conditions. The study also includes
developing the working design into an overall prototype. Testing is also performed to
validate the design as the most advantageous choice.
i
Acknowledgements
I would first like to thank my parents for supporting my studies (and sometimes
my wallet) and for always having a encouraging attitude and a patient hand. Without
them, I would never have accomplished what I have so far. A thanks to the rest of my
family for their encouragement. Also, a large THANK YOU to my girlfriend for her
support. Without her here, I would have gone more insane than I already have
I would also like to thank my committee for serving in their capacity. Special
thanks to my co-chairmen, first Dr. Walter O’Brien for his vast knowledge, willingness to
help and easy going attitude that made working for the department head less stressful
than I thought it would be. Thanks to Dr. Peter King, whose friendship, knowledge,
PATIENCE and large amounts of guidance enabled me to start and finish my graduate
studies.
The project support of Dave Beale and Jim Reed at Sverdrup Technologies was
greatly appreciated.
Thanks to those who came before me on this research: Tony, Julien, Grant and
Christian. Without their prior work, I would not have known where to start or what to do.
Finally, thanks to all those who endured my senseless rants and overall
randomness during my stay in the Turbolab. To those of old, Scott (Dr. Evil) for getting
me started, Grant and Christian for providing a never-ending source of entertainment and
Maj. Keith Boyer for providing me with someone to model my professional goals after.
To those I’ve met more recent: Mac, Jon, John, Rob, Melissa, Mike, Mono and Matthew
I thank for their friendship that made “The Hole” more tolerable. And a special thanks to
Joe (Mr. Evil) whose help was appreciated but whose patience, understanding and ability
to maintain his sanity in such close proximity to me were a miracle.
ii
Table of Contents
List of Figures…………………………………………………………....vi List of Equations ............................................................................ vii List of Tables ................................................................................... ix
1.1 Performance Effects of Distortion......................................................................... 1 1.2 Aeromechanical Effects of Distortion ................................................................... 2 1.3 Current Testing of Distortion ................................................................................ 3
2 Literature Review........................................................................... 5
2.1 Swirl Distortion ....................................................................................................... 6 2.2 Total Temperature Distortion ............................................................................... 7 2.3 Total Pressure Distortion and Performance Effects............................................ 9
Early Analysis ........................................................................................................................ 9 Distortion Analysis by the S-16 Committee......................................................................... 10 Modern Analysis .................................................................................................................. 12 Modeling .............................................................................................................................. 13
2.4 Total Pressure Distortion and Aeromechanical Effects .................................... 18 Early Analysis ...................................................................................................................... 19 Modern Analysis .................................................................................................................. 21 Modeling .............................................................................................................................. 21
2.5 Current Methodologies......................................................................................... 23 Tests of Distorted Flows ...................................................................................................... 23 Analysis of Distorted Flows................................................................................................. 25
2.6 Motivation for Work ............................................................................................ 28 2.7 Scope of Current Research .................................................................................. 30
List of Figures Figure 1-1: Graph Showing the Factors that Degrade the Surge Line and Distortion Surge Margin, (SAE, 1999) ................................................................................................ 2 Figure 1-2: Example of a serpentine inlet duct. (Small, 2001).......................................... 4 Figure 2-1: Timeline showing the approximate time from design to first flight. (Younghans and Paul, 1989)............................................................................................... 5 Figure 2-2: Exhaust Reingestion of a V/STOL Aircraft, (SAE, 1991)............................... 7 Figure 2-3: Typical Types of Distortion Inducing Screens (Eddy, 2001) ....................... 11 Figure 2-4: Comparison of Accuracy for Two Volterra Methods and the FRF (Luedke, 2001) ................................................................................................................................. 13 Figure 2-5: Pictorial Representation of Parallel Compressor Theory (Reid, 1969) ........ 14 Figure 2-6: Actual Flow’s Deviation from Predicted Square Wave Pressure Pattern, (Roberts et al., 1968)......................................................................................................... 15 Figure 2-7: Multiple Stream-Tube Model Versus Parallel Compressor Theory, (Mazzaway, 1968)............................................................................................................. 16 Figure 2-8: Comparison of Actuator Disk Predicted Pressures Versus Experimental, (Colpin and Kool, 1978) ................................................................................................... 17 Figure 2-9: The Different Types of Flutter and Their Placement on a Compressor Map, (Carta, 1989) ..................................................................................................................... 19 Figure 2-10: Relationship Between Distortion Intensity and Vibratory Effects, (Danforth, 1975) ................................................................................................................................. 20 Figure 2-11: TEACC Methodology, (Davis et al., 1998) ................................................ 22 Figure 2-12: Example of a Screen that Models Actual Flight Distortions, (Mokelke, 1974) ................................................................................................................................. 23 Figure 2-13: One, Two and Three-Per-Rev Screens......................................................... 24 Figure 2-14: Air-Jet Distortion Generator, (Overall, 1976)............................................. 25 Figure 2-15: S-16 Definition of Surge Margin, (ARP-1420, 1999) ................................ 26 Figure 2-16: S-16 Correlation Coefficients With Representative Screen Diagrams (Steenken, 1989) ............................................................................................................... 27 Figure 2-17: Typical Campbell Diagram With Critical Speeds Marked, (Manwaring, 1996) ................................................................................................................................. 28 Figure 3-1: Comparison of two wedges and their superposition, 3 inches behind wedge, for (a) aspect ratio less than 1 and (b) greater than one. Note that (a) is additive and (b) is not. .................................................................................................................................... 32 Figure 3-2: Total pressure drop, in percent of free-stream pressure, caused by 0o total angle wedge. Distortion due to the support rod is noted with white arrows.................. 33 Figure 3-3: Total pressure drop, in percent of free-stream pressure, caused by 15o total angle wedge. Distortion due to the support rod is noted with white arrows.................. 34 Figure 3-4: Comparison of stability limits for types of distortion. (Adapted from Davis et al., 2001) ........................................................................................................................... 35 Figure 3-5: Compressor map showing steady-state distortion stall (large points) and transient distortion stall (small points). Notice that the stall limit for transient approaches that for steady.................................................................................................................... 36 Figure 3-4: Pressure forces in Mach 0.6 flow for square wedges.................................... 38 Figure 3-5: Free-body diagram of ½ wedge with forces resulting from the flow. .......... 40
iv
Figure 3-6: Comparison of resultant force and its components for flow M=0.1 ............. 41 Figure 4-1: Drawing of front supported wedge. .............................................................. 43 Figure 4-2: Actuation forces for varying angles of a front-supported hinge. The force makes a large jump from 80o to 90o.................................................................................. 44 Figure 4-3: Drawing of rear supported hinge .................................................................. 45 Figure 4-4: Actuation forces for varying angles of a rear-supported wedge. Again, the force makes a large jump from 80o to 90o......................................................................... 46 Figure 4-5: Standard configuration, center-supported hinge. .......................................... 47 Figure 4-6: Vertical configuration, center-supported wedge........................................... 48 Figure 4-7: Actuation force for varying angles of a center-supported wedge. ................ 48 Figure 5.1: Lead screw design for actuation. The nut is (a) fixed and (b) movable. ...... 53 Figure 5.2: Piston-Cylinder design for actuation............................................................. 54 Figure 5-3: Electromagnetic design for actuation............................................................ 58 Figure 5-4: Piezo-strip design for actuation..................................................................... 60 Figure 6-1: Overall free body diagram for selected design. ............................................ 64 Figure 6-2: Diagram of overall distortion generator design ............................................ 66 Figure 6-3: Model of the selected distortion generator design ........................................ 68 Figure 7-1: Wind tunnel at Virginia Tech with major components listed. ...................... 70 Figure 7-2: Diagram of hinge used in distortion generator model (Adapted from Eddy, 2001). ................................................................................................................................ 71 Figure 7-3: Example of two individual wedges, connected together............................... 71 Figure 7-4: Side view of support structure with the sections cut out crosshatched......... 73 Figure 7-5: Distortion generator model a) Front view and b) Side view showing the slot............................................................................................................................................ 73 Figure 7-6: Test apparatus with the spring attached and the four test points labeled...... 74 Figure 7-7: Non-dimensional graph of predicted flow forces and measured actuation forces................................................................................................................................. 77 Figure 7-8: Picture of prototype connected to test cell. .................................................... 79 Figure 8-1: a) Diagram of final design and b) Picture of actual distortion generator...... 82 Figure 8-2: Example radial array of split airfoils making up the distortion generator. ... 83 Figure 8-3: Representation of distortion generator as a four-bar linkage. Specifically, a
slider-crank mechanism, with 1 being the crank and 4 being the slider……………85 Figure A-1: Drag forces in Mach 0.6 flow for multiple sizes of square wedges. ............. 86 Figure A-2: Lift forces in Mach 0.6 flow for multiples sizes of square wedges. ............. 86 Figure B-1: Pressure forces in Mach 0.1 flow for multiples sizes of square wedges. ...... 87 Figure B-2: Drag forces in Mach 0.1 flow for multiple sizes of square wedges. ............. 87 Figure B-3: Lift forces in Mach 0.1 flow for multiples sizes of square wedges............... 88 Figure C-1: Free-body diagram for front-supported wedge. ............................................ 89 Figure C-2: Free-body diagram for rear-supported wedge. .............................................. 89 Figure C-3: Free-body diagram for center-supported wedge. .......................................... 90 Figure D-1: Free-body diagram for rotationally actuated wedge. .................................... 91 Figure D-2: Free-body diagram for horizontally actuated wedge. ................................... 91 Figure D-3: Free-body diagram for vertically actuated wedge......................................... 92 Figure D-4: Free-body diagram for vertically-supported, actuated hinge. ....................... 92 Figure E-1: Force comparison for flow of Mach 0.107. ................................................... 93 Figure E-2: Force comparison for flow of Mach 0.103. ................................................... 93
v
Figure E-3: Force comparison for flow of Mach 0.1. ....................................................... 94 Figure E-4: Force comparison for flow of Mach 0.08. ..................................................... 94
In this graph the first and second bending (flexural) modes, the first torsional
mode and the second stripe (combination of bending and torsional) mode are shown with
the critical speeds associated with the specific engine tested indicated. This graph helps
the engineer to discover if aeromechanical problems exist in their design, because if a
natural frequency crosses an occurring per-rev line, and that speed is within the range of
operating speeds, a HCF problem will occur.
2.6 Motivation for Work
After reviewing the many different works encompassing the areas of analysis,
modeling and testing of distorted flows, it can be seen that there has been placed a great
deal of importance and interest on non-uniformities over the years. But one theme was
repeated throughout all of the past work, a heavy reliance on experimental data.
Therefore, researchers have explored more effective and efficient methods of
conducting tests. Currently, most test engineers use the direct connect screen method to
provide the necessary data. While this provides accurate distortion patterns, it does not
effectively model all situations, such as transient distortions. Currently transient
28
distortion phenomena are modeled with screens by simply testing the maximum
distortion levels observed in actual transient distortion occurrences. Screens are also very
inefficient, because in order to change the distortion pattern new screens must be
constructed, which can become very costly. In addition, labor increases because the
screens are employed through the direct connect method, meaning that in order to change
the distortion pattern engineers need to manually disconnect old screens and connect the
new patterned screens. The first method that attempted to fix labor and efficiency
problems was the airjet distortion generator. Its greatest drawback is that in introducing
air streams that change the total pressure and large support struts that block the flow, the
mass flow rate at the engine face can be reduced from that at the inlet.
In studying these current test methods, it is clear that a new method must be
developed to generate distortions. This is an opinion shared by many including DiPietro
(1996). He researched options for new distortion concepts, beginning with ten ideas
utilizing flow blockage (much as a screen does) and momentum exchange (as in an airjet
distortion generator). After studying the concepts, he found that because of the ease and
lower costs, the flow blockage method was more desirable. Also of advantage is that the
physical blockages can cause larger and more controllable distortion patterns. After
studying many geometries of blockage, DiPietro decided to implement a “split airfoil”
design by using a wedge shape to block the flow. As the angle of the wedge is changed,
it produces differing extents of distortion. A large advantage of the wedge flow is that it
has fixed separation points. This allows the streamlines of the flow to exist independent
of velocity, thereby having no critical Reynolds number. The shape is then the main
parameter in producing drag and pressure drops (DiPietro, 1996).
After seeing the advantages of a new wedge shaped distortion apparatus, the idea
to develop a testing device utilizing it arose. Jumel (1999) and Eddy (2001) studied the
distortions that were produced by a wedge in the airflow. Jumel used a static wedge in
order to determine the distance downstream of the wedge the distortion extended. He
found that the distortion patterns extended far enough downstream to be used to distort an
engine from a safe distance in front of that engine. Also discovered was that the patterns
and extents of distortion were easy to predict and therefore easy to accurately generate.
Eddy characterized the distortions left by multiple angle combinations for two wedges.
29
His tests proved that much like the results from Jumel’s single wedge tests, the distortion
due to multiple wedges was easily characterized. These studies lend credibility to the
wedge design. But will it meet the criteria set up in developing a pressure distortion
generator? Davis et al. (2001) put forth guidelines and requirements for this
development. He hopes to be able to meet seven areas that the next generation of
distortion generators must address:
1) Reduction of test cost and cycle time,
2) The evolution of advanced inlet systems,
3) The implementation of super-maneuverability,
4) In-flight weapon launches in aircraft featuring supercruise and stealth capabilities,
5) V/STOL aircraft operation in ground effect,
6) The advent of engines employing light-weight and highly-loaded compressor
stages, and
7) Engine performance enhancement through surge margin reduction or active stall
control.
In looking at these requirements it was found that the splitting airfoil design that DiPietro
proposed should be able to meet the criteria.
2.7 Scope of Current Research
The research performed was the initial design of the distortion generator and the
actuation device. The study’s purpose was to develop a concept that would be an
improvement on current distortion generation techniques. To that end, the driving goals
throughout the entire process were: 1) to meet the design requirements and 2) to
minimize the necessary actuation forces. Each area provided its own problems and
approaches. However, the methods used for selecting the best idea were the same for all
aspects. The design requirements were used as initial guidelines for developing many
possible versions. Then, the actuation forces were looked at in order to choose the best
concept for implementation.
30
3 Design Requirements
Studies done by DiPietro (1996) and Davis et al. (2001) not only provided general
ideas about distortion generator design, but also supplied specific criteria that a distortion
generator should have. These were condensed into actual design requirements by Dave
Beale of Sverdrup Technologies (Beale, 1997) assigning initial requirements that
included specifications for movement rate, size and angular movement. Also included
were requirements for maximum operating conditions of the surrounding environment.
As was discussed at the end of the Literature Review, a split airfoil design was
selected. To simplify the concept, flat plates were used in place of airfoils creating a
simple wedge. The advantages of a wedge design are rooted in its geometry, where the
overall moments from each inclined plate making up half of the wedge are canceled by
the other, symmetrical half-wedge. Also, if the boundary layer forces are assumed
negligible, with this assumption studied later, the distributed pressure force can be
modeled as a point force at the center of the half-wedge. The symmetry of the overall
wedge allows that each individual half-wedge moment will cancel and contribute nothing
to the opening and closing of the wedge. After selecting the underlying design, the
criteria for the distortion generator had to be defined in order to form more detailed
designs.
3.1 Size of Wedges
In order to gain acceptance of a new distortion generator within the engine testing
community, it must take the place of existing generators for all situations. Engines
ranging from small turboprops (GE CT-7 diameter of 25 inches) to large commercial hi-
bypass turbofans (GE-90 diameter of 158 inches) may require distortion testing and need
accommodating distortion generators. Currently, the most versatile method of creating
non-uniformities is the direct connect screen, which can be adapted to any type of engine,
thereby forcing a new generator to be just as adaptable. This means that a new testing
device must be scalable to different sizes. Studies done by the research contractors found
31
that a size range of 1 inch by 1 inch up to 5 inches by 5 inches overall wedge dimensions
would achieve the universality desired. This is an important parameter to define early in
the design process, because it will be a factor in determining many other requirements for
the device.
3.2 Range of Motion
Early on it was determined a wedge-type distortion generator would be required
to move from a fully closed position (0o) to a fully open position (180o). When the
required size range is considered, this motion requirement could mean as much as a 10-
inch travel. But, based on further studies, the range of angles the wedge must go through
can be reduced.
The first range restriction can be inferred from the work of Eddy (2001). He
studied the pressure distortion patterns of two wedges in both a horizontally aligned and
vertically aligned pattern. This study involved multiple angle combinations and found
that the overall pressure pattern behaves as a combination of the two individual patterns
for aspect ratios less than 1 but not at aspect ratios greater than 1, as seen in Figure 3-1a
and Figure 3-1b. The aspect ratio is defined as the ratio of the frontal height to the width
of the wedge, and therefore is equal to one at a total angle of 60o. Following this
suggestion in the design of a distortion generator allows the range to be greatly reduced.
a)
0 0.2 0.4 0.6 0.8
1 1.2 1.4
3 4 5 6 7 8 9 Horizontal Distance (i h )
Cp
(%)
30 and 90 30 90 b)
0 0.2 0.4 0.6 0.8
1 1.2 1.4 1.6
3 4 5 6 7 8 9 Horizontal Distance (i h )
Cp
(%)
90 and 120 90 120
Figure 3-1: Comparison of two wedges and their superposition, 3 inches behind wedge, for (a) aspect ratio less than 1 and (b) greater than one. Note that (a) is additive and (b) is not.
The next study was performed to determine the minimum angle the wedge could
be closed to without creating a notable distortion in the flow. Studies into the magnitude
32
and extent of distortion were performed at wedge total angles of 0o and 15o, with the
results shown in Figure 3-2. The results observed from the closed wedge were reported
as a total pressure coefficient, defined in equation 3.1,
100*ambo
iop PP
PPC
+−
= (3.1)
where, Po-Pi is the pressure difference measured, Po is the inlet total pressure and Pamb is
the ambient pressure. This coefficient had a drop in maximum magnitude of 0.4-0.5
percent and an overall area (neglecting the support rod denoted by the arrows) of pressure
distortion measuring 0.8125 square inches. These results were then compared to those of
the wedge at an angle of 15o, and as shown in Figure 3-3, the 15o wedge results in a
maximum drop in total pressure coefficient of 0.4-0.5 percent as well. Differences from
the closed wedge appear when the area of pressure drop is considered. From Figure 3-3,
it can be seen that the distorted area of the 15o wedge is 1 square inch. This is an
approximately 20% increase in distorted area and mandates that the wedge must be
closed to 0o in order to maintain a minimal disturbance to the flow.
4.75
5.25
5.75
6.25
6.75
7.25
7.75
6.754.7555.255.55.7566.256.56.75
Horz. Position (in.)
Vert. Position (in.)
0.4-0.50.3-0.40.2-0.30.1-0.20-0.1
Figure 3-2: Total pressure drop, in percent of free-stream pressure, caused by 0o total angle wedge. Distortion due to the support rod is noted with white arrows.
33
4.75
5.25
5.75
6.25
6.75
7.25
7.75
8.25
4.7555.255.55.7566.256.56.75
Horz. Position (in.)
Vert. Position (in.)
0.4-0.50.3-0.40.2-0.30.1-0.20-0.1
Figure 3-3: Total pressure drop, in percent of free-stream pressure, caused by 15o total angle wedge. Distortion due to the support rod is noted with white arrows.
From the above studies, the operating range of motion for the wedge is a
minimum angle of 0o and a maximum opening angle of 60o. These studies greatly reduce
the amount of travel necessary for the actuation mechanism, which can lead to smaller
and less expensive actuators to meet the requirements.
3.3 Rate of Actuation
Transient effects of distortion have been a growing area of importance, starting
with the experiments performed on the F-111 aircraft mentioned in the section entitled
“Total Pressure Distortion and Performance Effects.” A new distortion generator will
need to not only replace existing methods, but improve on them as well. To this end, a
new distortion generator will include the capability to model time-dependant distortion
effects as well as steady-state effects.
Currently, test engineers use the peak distortion level over a representative period
of time to model transient effects, such as turbulence. This method results in over-
design, by creating an engine that can operate under the worst distortion levels all of the
time. But this does not take into account the time spent at peak levels or the vibratory
effects of constantly changing pressure levels. Other problems with the representative
peak level method include the limitation that this test only accounts for transient
magnitudes, and not positions.
34
The most common problem involving transient positions is inlet swirl.
Directional transients such as swirl can reduce the stability limit of an engine. An
example of this is shown in Figure 3.4, where stability limits are shown for a clean inlet,
stationary distortion and swirl distortion (both co and counter rotation). Also involved
with swirl distortion is the inherent vibratory problems with a single blade being
subjected to the cyclic nature of swirling flows.
Figure 3-4: Comparison of stability limits for types of distortion. (Adapted from Davis et al., 2001)
Distortion position and magnitude transients can be modeled by changing the
intensity and extent of distortion, and then measuring pressures over this entire process.
The most pressing problem with modeling transient processes is how to achieve realistic
patterns and changing flows. This problem creates a question of how fast a modeling
device should change the distortion in order to accurately model distortion occurrences.
The frequency of transient variations that should be studied is not known, but the
current contract requires full travel in 0.5 seconds. If the maximum travel is considered
(10 inches, 180o), the requirement becomes a 20-inch per second (360 degree per second)
rate. But, if the range restrictions evaluated earlier are imposed on the design, the
maximum rate reduces to 5-inches per second (120 degrees per second). It is important
to note that lower rates (1.5 degrees per second) do not introduce performance problems
seen elsewhere, as can be seen in Figure 3-5. Therefore, studies involving transient
distortion should test at higher rates, with the current design having a rate requirement of
120 degrees per second, used in order to meet the contractor’s requirements after
assessing restrictions on the total travel.
35
Figure 3-5: Compressor map showing steady-state distortion stall (large points) and transient distortion stall (small points). Notice that the stall limit for transient approaches that for steady.
3.4 Test Cell Conditions
An important factor in any design process is defining the operation conditions.
The volatile nature of turbomachinery introduces uncommonly high conditions. In order
to develop a test device that will be used in many different situations, the operating
conditions must cover a large range. Given as the upper conditions by the contractors are
a Mach number of 0.6, a free-stream total pressure of 40 psia, and a free-stream total
temperature of 300oF. These conditions are important in material selection and device
structure, because the materials must not degrade under the high pressures and
temperatures and the device must be able to withstand the drag forces placed on it by the
high air speeds.
3.5 Flow Conditions and Forces
In order to properly size the support structure as well as the actuation system, the
forces acting on the system need to be determined. To assure that the distortion generator
will operate correctly under the range of operating conditions given above, the forces
were calculated for the largest values of test cell conditions for a range of wedge sizes.
36
Forces on the airfoil are generated by a pressure difference between the upper and lower
surface. The resulting pressure force, Fpress, acts perpendicular to the plate. The pressure
force can be calculated using,
2
2 AUCF Npressρ
= (3.2)
where, CN is the normal coefficient, ρ is the density, U is the free stream velocity and A
is the reference area. The normal force coefficient is a function of the angle of the plate,
and is given for multiple angles in Table 3.1. These values were determined from
experimental data on a three-dimensional inclined plate representing half of the wedge
design, found in Blevins (1984).
Table 3-1: Normal force coefficients for varying angles (adapted from Blevins, 1984)
where, the variables are defined in the “Actuation Design” section. The rotational
direction remained a torque because, the aspect being compared was the dependence on
the flow and not the computational values that would require a conversion to force. As
can be seen from these equations, the vertical-supported direction is independent of the
flow forces and was, therefore, selected. Actuation forces that are independent of the
flow are important in the design because it allows the same force to be used at all flow
speeds and all wedge sizes. The support force seen in equation 5.4 is directly related to
the actuation force, meaning that this value is a function of the magnitude of force
applied to the wedge for actuation. This is a complicated problem, but it is not important
for design selection. The important conclusion from this comparison is that the
vertically-supported structure has an actuation force independent of the flow.
Also, when comparing the two center supported designs to each other, the
vertically configured method is simpler. The support structure running in the same
direction as the actuation travel allows easier transmission of motion to the outside, if that
becomes necessary. The standard configuration would require a 90-degree change of
direction, assuming the actuation motion enters parallel to the support structure. Another
advantage of the vertical configuration is that any linear actuation can be contained inside
the support structure, reducing the FOD hazard. This design configuration will also not
require a support to maintain its orientation to the flow, as others would.
Implementing these selection criteria allowed us to choose the wedge design: a
center supported, vertical orientation wedge with vertically supported actuation direction.
This overall selection will provide the lowest overall needed actuation force. The
necessary overall force is only the support force, Fsupy . The FBD for the combined
support forces and actuation direction is shown in Figure 6-1. The weights are included
here in order to calculate the overall actuation force necessary (Foverall). This also was
63
chosen as the best concept because of the many other advantages found with the
individual evaluations of wedge design and actuation direction.
F p
F D
F L
F p F L
F D
F sup y F sup x
F sup x F overall
x y
Q
W 1
W 1
W 2
Flow
F overall = F supy + (2*W 1 ) + W 2
Figure 6-1: Overall free body diagram for selected design.
6.2 Actuator Selection
As was the case with the wedge selection, actuation selection was done by
looking at multiple characteristics independently. After determining the best design for
each aspect of actuation, they were combined to form the final actuation design. The
aspects looked at were actuation direction, type of actuation and actuation placement.
Some of these features also had additional characteristics that had to be decided upon.
The direction of actuation was already chosen during wedge selection. Therefore,
the selection process will not be repeated, only the result. The vertically supported
direction was decided upon for its low necessary actuation forces.
Next, the type of actuation was studied, with the initial concepts for moving the
wedge presented in the “Actuation Design” section. Before comparing the designs,
selection criteria were developed. These criteria were applied to the designs in order of
importance. The first design condition was the necessary size to force ratio. This
determined if the actuator generated the force needed for actuation while maintaining a
small size. The concepts that passed this standard were the motor, pneumatic cylinder
64
and hydraulic cylinder. The next criterion was a comparison the overall number and
complexity of the system’s components. This was important to provide the actuator with
an ease of operation. The complexity of hydraulics proved it a poor choice for passing
this standard. A hydraulic system has many components including a pump, accumulator,
and specialized hoses. Finally, the selection process for the type of actuation investigated
the simplicity of the overall system and the FOD hazard. These are included together
because with a structurally simple design comes a low FOD hazard. Here, the motor
passed on because of the method of force generation a pneumatic device uses. The air
required to move a pneumatic piston must be pressurized. With this added pressure
comes the hazard of rupturing the containment structure. Also, as was mentioned in the
“Actuation Design” section, a pneumatic piston requires a valve close to the movement
point. This could introduce a FOD hazard if an inside positioning scheme were used.
The final actuation design aspect investigated was the placement of the actuator.
For the requirements set forth, an outside placement was chosen. This was done because
of the low flow disruption, lack of size restrictions, lack of operating condition
restrictions and low FOD hazard associated with placing the actuator outside of the test
section. Additionally, the overall effectiveness of the distortion generator will improve
with the ability of the wedge to completely close, which can only be accomplished by
placing the actuator outside of the test cell. Also, ease of both operation and maintenance
are increased with the lack of size and operating condition restrictions. Less hardware
restrictions can also prove to be a more cost effective scheme by allowing more choices
in the models of actuator to be used. Finally, with fewer moving parts inside the test
section, the risk of a FOD incident is drastically reduced.
Based on the assessments of the three most important aspects of the actuator, the
final design chosen was a motor actuating in the vertically supported configuration,
placed outside of the test cell. With this decided, specifics of the design had to be
decided upon, with the first to be the type of motor that would be used. A stepper motor
was chosen because of the simplicity of its open loop control scheme and positive
positioning aspect. This will make the design of a control system as well as the operation
of the distortion generator less complicated for those in the future. Also, it was decided
to use a combination of a lead screw and rods to transfer the force to the wedge inside the
65
test section. A lead screw would be used to convert the rotational motion of the motor to
a linear motion. The rod would be used to then transfer this linear motion from outside
the test section to the wedge. This configuration would eliminate the potential problems
associated with adding a rotational movement to the linearly moving components
necessary inside the test section. A lead screw was chosen over a rack and pinion set-up
because of the lead screw’s ability to be used in any orientation. A rack and pinion can
only be used in horizontal configurations without any added support.
6.3 Overall Design
Based on the above two selections, the overall design chosen was a vertical
wedge, supported in the center and actuated by a motor driving a lead screw. A diagram
of this configuration is shown in Figure 6-2. This design will provide the optimum
minimization of the necessary actuation force and a method of actuation that will work
under all of the design requirements. The next step was to work out the design specifics
of the distortion generator. This was done by constructing a concept model for design, as
opposed to operational, purposes.
����������������
������
Stepper Motor
Lead Screw
Traveling Nut
Vertical Connection Rod
Support Structure
Distortion Generator
Horizontal Support Rod
Figure 6-2: Diagram of overall distortion generator design
66
6.4 Concept Model Fabrication
A model, seen in Figure 6-3, was built of the selected distortion generator in order
to identify methods of construction and problems that may occur. This model was a
movable representation of a three-part array using the selected design. The first
information gathered in the construction of the model was that the linking of the vertical
connection rod and the horizontal support rod required a 90o connection. This was solved
by using a small eyebolt on the end of the vertical connection rod to move the horizontal
support rod. Also, the slot that was cut into the wedge to accommodate the support
structure needed to be larger than originally thought. In order to allow for a fully open
position, the slot had to extend the full length of the wedge. This was solved by deciding
that two individual wedges would be connected together by the horizontal support rod.
This would allow the length of the slot to extend to the ends of the wedge, and the width
of the slot to be determined during construction so it would allow enough room for the
support structure to not interfere with movement. Another problem discovered during
construction of the model was one that cannot be fixed, but was inherent to the design.
Because one side of the wedge was fixed and the other could move in the vertical
direction, the center of the wedge moved vertically as the angle of the wedge changed.
The area of lower pressure will also shift as the distortion is created, causing patterns of
distortion that vary from the predicted. It is believed that this can be overcome by
compensating for the shift in a computerized control scheme. Constructing the model
also helped in the discovery of a design improvement. The support structure, originally
intended to be circular, was chosen as an airfoil shape. This improvement allowed less
flow disruption and more internal area for the vertical connection rods. This model was
very helpful as a first step in constructing the distortion generator. It allowed the layout
to be determined using materials that were inexpensive and easy to shape. Creating a
model was very beneficial as a precursor to the actual construction of prototype.
67
Figure 6-3: Model of the selected distortion generator design
68
7 Design Verification
After the design was selected, a method had to be developed to verify the
assertions made. The main focus of the testing was to show that the actuation forces
were low enough for practical purposes and were independent of the flow. In order for
this to be accomplished, a testing set-up had to be created that would allow the linear
actuation force outside of the test section to be measured.
7.1 Wind Tunnel and Test Cell
In order to measure the actuation forces necessary, the first task was to find an
acceptable wind tunnel for the testing. A tunnel built by a Virginia Tech student, Tkacik,
for his work on stalled flows was found to be a suitable choice, and is shown in
Figure 7-1 (Tkacik, 1982). The main reason this tunnel was selected was because of its
prior use in both Jumel’s and Eddy’s studies. In order to maintain a commonality with
these earlier studies, the same wind tunnel was selected for the testing. This wind tunnel
uses a model BIA, size 630 class II centrifugal blower built by Aerovent Corporation.
The fan is driven by a 15 horsepower electric motor and has an adjustable inlet area for
altering the flow speed. This is accomplished by variable inlet guide vanes on the fan’s
inlet. Original design speeds of 49.2 ft/s to 157.5 ft/s were reported, but for the testing of
this experiment, one speed was used, 114.4 ft/s. A honeycomb and 3 screens are located
inside the settling section in order to reduce free stream turbulence before entering the
nozzle. The flow speed is increased by a nozzle constriction from 9 ft2 to 1 ft2, which is
the exit area of the tunnel.
69
Blower
Settling SectionNozzle
Figure 7-1: Wind tunnel at Virginia Tech with major components listed.
A test section was constructed that would match this tunnel exit area. Eddy
constructed a 1 ft2, ½-inch thick Lexan™ test section. This test section was 3 1/3 feet
long, and had a mounting flange on one side in order to mate with the tunnel. This test
section, with some modifications, was determined suitable for the proposed test forces.
These modifications included cutting airfoil-shaped holes ½-inch long and ¼-inch wide,
aligned on both the top and bottom of the test section. These holes were made so the
support structure could protrude from the test section, allowing it to be fixed in place.
Adjacent to the airfoil shaped holes, slots were cut to allow for protruding supports from
the support strut that would hold the entire distortion generator in place during testing.
Also, a 4 in2 access hatch was cut into the top of the test section so that the distortion
generator could be constructed outside of the test section, and placed inside after it’s
completion. Finally, a 2/5-inch hole was drilled 4 inches in front of where the distortion
generator was placed. This hole allowed access to the free-stream flow for the pitot-static
probe in order to make velocity measurements. These modifications allowed the
distortion generator to be incorporated into the test cell for the force testing.
70
7.2 Distortion Generator
Although the design and layout of the distortion generator were defined in the
“Design Selection” section, some construction requirements were still needed. The most
important of these was the size of the wedge to be used. The “Design Requirements”
section lists the size range as 1 in2 to 25 in2. In order to closely model the tests of Jumel
and Eddy and generate forces significant enough to be measured, a wedge size of 2
inches by 2 inches was chosen. This was constructed from two, 2 inches tall by 1 inch
wide hinges. These were modified versions of the same model used by Eddy, McMaster-
Carr model 1624A51, as shown in Figure 7-2. Although the design called for a single
wedge with a slot cut down the middle for the support structure, two of the 2 by 1 inch
wedges were used and connected by horizontal support bars. A picture of this completed
generator element is shown in Figure 7-3.
1.00
2.00 0.25
Figure 7-2: Diagram of hinge used in distortion generator model (Adapted from Eddy, 2001).
Horizontal port
Horizontal port
Figure 7-3: Example of two individual wedg
These size of these support bars, as well as the
The supports examined were the horizontal support ro
71
SupSupRods
es, connected together.
other supports, had to be selected.
ds, the vertical connection rod and
the overall support structure. The horizontal rods had to be able to withstand the bending
forces placed on them, which was the drag force generated by the wedge. The maximum
drag force that would be encountered would be 1 lbf, for a 4 in2 wedge in a Mach 0.1
flow. The horizontal support rod was modeled as a simply supported beam, where the
maximum deflection was given in Beer and Johnston (1992) as,
EI
FL48
3
=δ (7-1)
where: δ = deflection F = force on beam L = length of beam E = modulus of elasticity
I = moment of inertia
A 1/16-inch diameter, 2-inch-long piece of steel music wire was chosen for its size
relationship to the overall generator. This would experience a maximum deflection,
under the above-mentioned conditions, of 75 mils. This is more than acceptable for the
tests to be conducted.
The next support studied was the overall support structure (airfoil). Again it was
modeled as a simply supported beam, but this time the shape was approximated as a
hollow, aluminum rectangle. The calculated deflection for this support was found to be
small and definitely not of concern.
Finally, the vertical connection rod was not examined for bending. The only
forces acting on this support are axial ones, since it is not exposed to the flow. The steel
music wire used (3/32-inch diameter) was assumed to be strong enough for the axial forces
being used.
The next step in constructing the distortion generator was to cut holes into the
support structure for the horizontal support rods. The selected support was a 16 inches
long by ¼-inch wide, ½-inch chord, 1/100-inch thick aluminum airfoil. The upper pin had a
slot manufactured instead of a hole so that when the vertical support rod was connected
to the horizontal support rod, the horizontal one could move through the range of motion
of the wedge. The drawing used for modifying the support structure is shown in Figure
7-4. The clearances between the cuts and the horizontal support rods were kept small
(1/64-inch) in order to minimize the horizontal vibrational motion of the overall wedge in
the airflow.
72
Figure 7-4: Side view of support structure with the sections cut out crosshatched.
The final step in construction of the distortion generator was the assembly. The
vertical support rod is connected to the top horizontal support rod by using an eyelet
connector. This piece of hardware has a connection end diameter of 3/32-inch and an eye
diameter of 1/16-inch. The horizontal supports were connected to the wedges using a
quick drying, metal epoxy. Graphite powder was spread on the slot and hole to reduce
the friction generated during movement. The assembled generator can be seen in Figure
7-5, where it is important to note that it is assembled upside down from Figure 7-4. This
was done so that if a failure occurred, the wedge would close to the 0o position.
a) b)
Figure 7-5: Distortion generator model a) Front view and b) Side view showing the slot.
73
7.3 Testing Apparatus
In order for the design of the newly created distortion generator to be acceptable,
its must act in a method very close to the predicted manner. A device needed to be
created so that it could be determined if the actual device was following the predicted
behavior. The primary behavior that was studied in order to validate the distortion
generator’s performance was the actuation force. This was the main driving factor in the
design and remained the main driving factor in testing as well. Therefore, a test
mechanism had to be developed to measure the actuation force at pre-determined
operating conditions.
It was decided that the simplest method of measuring the actuation forces for
different conditions was by using a spring. A spring will move a specified distance for a
given force, with this distance dependant on the spring constant, being 0.2 pounds per
inch for this specific spring. Therefore, a test stand was constructed to provide a
stationary attachment point for one end of the spring, while the other end was connected
to the vertical support rod. In this design, as more force was applied to the wedge from
the flow, the spring stretched to a larger distance. If this distance was measured, the
force applied could be calculated. The test apparatus can be seen in Figure 7-6 with a
ruler attached for measuring the change in spring length.
Figure 7-6: Test apparatus with the spring attached and the four test points labeled.
74
To see if the performance of the device followed predictions, the force was
measured at four different starting angles for a flow speed of Mach 0.1. For this to be
accomplished, the apparatus was developed to measure the force at four arbitrarily
chosen angles. The angles were set by moving the stationary spring connection to one of
the four points shown above in Figure 7-6. By moving the stationary point, the starting
angle (i.e. no flow) of the wedge was changed. This allowed different amounts of drag
and lift forces to be generated based on the angle. Therefore, since the actuation force is
dependent on the flow forces, the test apparatus would measure more force as the initial
angle was increased (by increasing the number slot the stationary end of the spring was
placed).
7.4 Test Results
The tests were performed at the four, previously mentioned spring positions four
separate times, for repeatability. These positions correspond to starting angles of 45o,
75o, 77o, and 160o. The four tests were conducted at approximately the same speed,
M=0.1 with an Mach number error range of 0.027. Because this was a low-force design,
the desired speed chosen was at the higher end of the wind tunnel’s operating range. In
theory, this higher speed would create larger actuation forces that would be more readily
measurable by the test apparatus.
The data in Table 7-1 gives the final angles and forces measured in the testing. It
is important to note that during the calculation of the forces, the weight of the wedge and
the initial resistance of the spring had to be added to the force calculated from spring
movement having values of 0.05 pounds and 0.09 pounds, respectively.
75
Table 7-1: Angles and forces for tested speeds
Mach Number Angle (degrees) Force (lbf)
0.107 30 0.0527
0.107 54 0.0527
0.107 62 0.0652
0.107 Inverted N/A
0.103 32 0.0527
0.103 59 0.0527
0.103 64 0.0652
0.103 Inverted N/A
0.100 34 0.0527
0.100 62 0.0527
0.100 65 0.0527
0.100 Inverted N/A
0.080 36 0.0527
0.080 69 0.0527
0.080 78 0.0527
0.080 Inverted N/A
As can be seen from these results, the necessary actuation forces are very small.
The “Inverted” angles were starting angles in which the flow force aided in opening the
wedge. The test apparatus had no stopping mechanism for this direction because it only
had a spring attached on one side of the vertical connection rod. Therefore, the wedge
continued to open until it turned itself inside out. This was expected to happen at any
total angle in excess of 90o. The test points where the test angle was inverted were
discarded in data analysis. This is acceptable because, as described in the “Design
Requirements” section, the range of travel is only necessary up to a total angle of 60o.
Further reduction of the data was needed in order to gain insightful knowledge into the
system.
For comparison of the measured actuation force to the calculated flow forces, a
non-dimensional approach was determined in order to negate the effects of the speed
variations. Because the predicted flow forces were all calculated from published lift and
76
drag coefficients, the tested actuation force was non-dimensionalized in the same manner.
The equation is given as
22
1 VF
F ρ=C (7.2)
where, CF is a generic term referring to any force coefficient, F is the force, ρ is the
density and V is the velocity. This graph can be seen in Figure 7-7, with the dimensional
graphs given in Appendix E. The non-dimensional graph shows that the measured
actuation force is very small, compared to the flow forces. The actuation force is actually
lower than all forces, even the viscous forces. This confirms the earlier assumption that
the viscous forces can be neglected because of the symmetry of the wedge. It also
validates the design in that the actuation forces are not caused by the flow, but by other
factors such as weight.
0
0.0005
0.001
0.0015
0.002
0.0025
0.003
0 50 100 150 200
Total Angle (degrees)
Forc
e C
oeffi
cien
t
dragliftpressure viscous actuation
Figure 7-7: Non-dimensional graph of predicted flow forces and measured actuation forces.
7.5 Design Prototype
The next step in verifying that our design was a good choice for future
development involved the construction of a prototype distortion generator. This would
allow the stepper motor, as well as other components, to be sized. By using the results
from the analytical designs and the testing, a prototype was constructed that demonstrated
the device’s operation.
77
The first components sized were the lead screw and motor. These had to be
determined at the same time because the actuation force is generated by the combination
of stepper motor and lead screw. A combination of motor and lead screw had to be found
that could generate enough force to move the wedge. After looking at the test results, it
was determined that a force of 0.053 pounds would move the wedge under most
conditions (i.e. it would only need to overcome the weight and friction). Therefore, the
stepper motor could be chosen arbitrarily and the lead screw was chosen to transfer the
torque into the required linear actuation force. Throughout the selection process, size
relationships to the 4 in2 wedge were kept in mind so that construction would not be
overly complicated and to minimize the weight. The stepper motor chosen was a Nippon
Pulse Motor Co., Model PF55-48. This motor generates a maximum holding torque of
1.062 lbf-inch, rotates at a maximum speed of 387.5 RPM, and has a body diameter of
2.7 inches. Controlling the motor involved generating a simple pulse train using a
LabVIEW™ program, a National Instruments DAQCard™-AI-16E-4, and an
Alltronics™ model 5804-stepper motor driver chip. A lead screw having a diameter of 3/8-inch and a 50 percent efficiency (which is common for a lead screw and plastic nut)
generates, in conjunction with the above-mentioned motor, an actuation force of 0.1
pounds. Choosing this lead screw generated more force than was necessary, but was
used in order to allow for error in the efficiency. It is common for efficiencies of a lead
screw to be as low as 30 percent, which in this case would still generate sufficient force
(0.06 pounds). This lead screw was a McMaster-Carr model 99030A303, having 2
revolutions per inch (which when combined with the motor equals 3.23 inches per
second, or a minimum of 370.3 degrees per second). This rate was found by using the
travel of the lead screw/motor combination and the Law of Cosines to determine the
relationship of linear travel to the angle of the wedge. As a side consideration, this rate
was checked against that in the “Design Requirements” and found to be well in excess of
the 120 degrees per second goal.
The next component selected was the traveling nut that moves linearly along the
lead screw. The internal diameter of this nut is set by the size and pitch of the lead screw.
But, the material still had to be determined. By looking at various sources, it was
determined that a plastic nut would have a higher efficiency than a bronze or steel nut.
78
Therefore, the nut used was McMaster-Carr model 1349K101. This nut was made of
polyethylene terephthalate (PET), had an outer diameter of 1.35 inches and matched the
chosen lead screw. As mentioned before, this should provide an efficiency around 50
percent. Selecting the nut completed the power transmission element of the prototype.
Other components had to be constructed or altered as well, mostly for connection
purposes.
A coupling to connect the lead screw to the stepper motor was needed. It was
found that commercial rod connectors were too bulky and heavy for this application.
Therefore, a 5/32-inch hole was drilled in the center of the lead screw, and the shaft of the
stepper motor was secured in place with epoxy. Three 3/32-inch holes were drilled at 120o
intervals and 3/10-inch from the center. Finally, another 1/8-inch diameter setscrew hole
was drilled into the side of the nut, into one of the 3/32-inch holes, so that the vertical
support rod could be locked into the hole. The other three holes had guide rods running
through them. A frame was constructed to hold the actuation mechanism in the correct
orientation to the distortion generator. A picture of the completed prototype is shown in
Figure 7-8. This prototype was placed in a Mach 0.1 flow to prove that the concept
would work under operating conditions.
Leadscrew
Travelling Nut
Distortion Generator Support Structure
Figure 7-8: Picture of prototype connected to test cell.
79
8 Conclusions and Recommendations
8.1 Review
For more than forty years, gas turbine manufacturers have perceived a need to test
inlet distortion and its effects. The earliest studies concentrated on performance
evaluations, while newer studies also focus on structural integrity of the blades and high-
cycle fatigue. Some of the more important distortion-creating events include non-
uniformities in total temperature and swirl velocities. By far the most widely investigated
distortion causing event is the introduction of a total pressure drop. Current research
concentrates on steady-state tests to characterize an engine’s response to inlet distortion.
These tests methods often involve screens to produce the distortion. But current testing is
evolving to areas where screens are not applicable. Their usage is becoming too
inefficient and they are not testing all of the desired aspects, such as transient responses.
A push to replace screens came with the introduction of the Airjet Distortion Generator
by Overall (1977). This device used jets of air injected at an opposing direction to the
flow. This allowed the distortion patterns to be changed quickly as well as some degree
of transient studies to be performed. But, this device had drawbacks that limited its
usage. The transient tests could only be performed for short periods of time, because of
the short pulses that could be sent to the flow from the airjets. Also, this system
introduced blockages to the flow, in the form of support structures, which created
distortions in undesirable positions. The inadequacy of the Airjet Distortion Generator
and the inefficiency of the screens led to studies on a new design for a total pressure
distortion generator.
One idea for a new design that emerged was the concept of using a splitting airfoil
to generate patterns of distortion. DiPietro (1996) performed the preliminary studies on
this concept, finding it to be a promising idea. Jumel (1999) and Eddy (2001) further
studied this idea in order to define its operation and feasibility. They performed studies
that evaluated the pressure loss induced by the splitting airfoil design. Their studies
determined the concept to have exceptional promise in both the areas of efficiency and
transient testing. They also helped in the design of a system by characterizing the
80
pressure distortion that is created by a variable angle, splitting airfoil. The positive
results of all of these studies led the way to the design of an actual splitting airfoil
distortion generator.
8.2 Conclusions
In designing the distortion generator, the first aspects looked at were the system
requirements. The requirements for the design were set using the conclusions of Eddy,
results of previous transient studies, and recommendations from the Arnold Engineering
Development Center. The final requirements for the system were a size range of 1 in2 to
25 in2, a travel range of 0o to 60o, and a rate of 120o per second. Also, for simplification
purposes, a hinged wedge was used to represent a split airfoil. It acted in the same
manner an airfoil would, but the flow forces generated were greater. These requirements
were kept in mind in designing the system to be used for the distortion generator.
The next step in the design involved selecting the wedge configuration, with
respect to its support structure. Four designs were studied, with the result that some
designs allowed the support structure to absorb some of the flow forces. This reduced the
necessary actuation force for the generator. Also looked at were the actuation directions
that were possible for each wedge configuration. This was done in order to define the
forces that an actuation device would have to overcome. The design selected was the
vertically oriented, center-supported wedge with a vertical-supported actuation direction.
This design used the support structure to completely counteract the drag forces generated
by the flow, and required actuation at only one point.
By selecting the wedge configuration, the moving parts and in what direction they
moved were defined. How this actuation was to be performed was still to be determined.
A study of possible motion devices was performed, with a stepper motor chosen because
of the variations of force and size available, as well as its ease of control. The most
important factor in choosing the stepper motor is the positive placement aspect. This
means that no position feedback is necessary to move to specific angles. The method
chosen for converting the rotary motion of a motor to a linear motion was a combination
of a lead screw and a push/pull rod. This provided a system that has a good mechanical
81
efficiency (approximately 50%) while being able to operate in any orientation to the test
cell. This is important for when an array is implemented, as mentioned later.
Therefore, the final design combined the vertically oriented, center-supported
wedge, using a stepper motor in conjunction with a lead screw and a push/pull rod to
provide motion. A diagram and picture are shown in Figure 8-1.
a)
��������������������������
Stepper Motor
Lead Screw
Traveling Nut
Vertical Connection Rod
Support Structure
Distortion Generator
Horizontal Support Rod
b)
Figure 8-1: a) Diagram of final design and b) Picture of actual distortion generator.
Tests involving the constructed generator confirm our design choices. The
actuation force needed was extremely low and was not generated from the flow forces.
Also tested was the stepper motor’s feasibility in this design. This motor moved the
wedge to a fully open position in 0.637 seconds. While this is outside of the design
requirements, it is close for the generic motor used. If a better-matched motor was
chosen, the time requirements could be met. This time was larger than the predicted time
because of unaccounted for system efficiencies. If the travel restriction of a maximum
wedge angle equal to 60o, then this distortion generator meets the time requirement.
Therefore, a stepper motor will work as the actuation device for this design, and it
should be easy to scale, since large steppers will not be required for subsonic flow. This
configuration of the splitting airfoil will be an effective method of replacing screens for
distortion testing in aircraft engines. It will increase the efficiency in testing, by allowing
multiple conditions to be tested with minimal labor. This design is also very flexible in
its implementation because the necessary actuation force is independent of the flow and
82
allows the same actuation system to be used for various flow speeds. It will also allow a
new area of testing to be explored, transient effects, which is becoming more and more
important.
8.3 Recommendations
Although the final design was selected as the best choice, many areas of operation
still need to be studied. The most important aspect still to be studied is how this design
will operate in an array. The distortion patterns that can be created by an array need to be
studied in order to make sure they are acceptable. An example, radial array is shown in
Figure 8-2. For this configuration, the split airfoils could also be tapered to better fit the
circular geometry of an engine intake.
Figure 8-2: Example radial array of split airfoils making up the distortion generator.
The other primary area to be included in future studies is the control system.
Most likely, a computer will be used for the controller. But the design’s geometry leads
to some minor control problems. The wedge angle can be controlled by a simple
relationship to the linear distance traveled by the actuation device. But, the center of the
wedge will also move linearly with the actuation device. This does not change the
relationship of the distortion strength to the angle, but it does change the position of the
center of the distortion pattern. This could potentially shift the reduced pressure area, and
cause undesired patterns of distortion. Because this is an aspect of the design, the
system’s control should take into account the possible shifting of the distorted area. As
the number of generator elements increases, and the size of each decreases, this problem
is minimized. By compensating for this and having a number of airfoils suitable for the
necessary precision, predictable distortion patterns can still be created.
83
Also recommended is using a computational method to define the flow over the
wedge. This could provide important information regarding the viscous forces and 3-D
effects. It could determine if the actuation forces seen are affected at all by the viscous
forces, separations or three-dimensional effects.
Other possible causes of the forces that should be studied include friction and
construction imperfections. If the horizontal support rods are not exactly centered on the
plates of the wedge, then the moments generated on each side of the rod due to the flow
will not cancel each other out. This is one reason as to why the wedge has a tendency to
close at lower angles, and open at higher angles. If the larger moment is above the hinge
point, the wedge will tend to close. At higher angles, the moment can move below the
hinge point, attempting to open the wedge.
A kinematic approach could also be used to study the overall forces on the
system. The wedge acts as a four-bar linkage, with the type being classified as a slider-
crank mechanism. The mechanism is shown with the individual linkages numbered in
Figure 8-3. It is important to note that the design should take into account (if necessary)
the dead points located at the extreme extensions of a slider-crank. Kinematics studies of
four-bar linkages use the Law of Cosines to accurately predict the position of the
linkages, an idea that supports the earlier suggestion of using this law to correlate the
wedge angle to the linear travel. Kinematics could also provide a method to optimize the
size of the support structure (linkages 3 and 4) with respect to the wedge (linkages 1 and
2) and another method to determine axial forces in all of the members, and therefore the
actuation force necessary on linkage 4. It is recommended that in addition to a fluid
optimization, a kinematic optimization be performed on the mechanism.
������
���������
����������������������������
1
2
3
4
Figure 8-3: Representation of distortion generator as a four-bar linkage. Specifically, a slider-crank mechanism, with 1 being the crank and 4 being the slider.
84
The final recommendations for continuing the design of this distortion generator
include a study into the transient responses created by an array of split airfoils. This
could become important in characterizing the generator’s usefulness for transient testing.
The new design for the total pressure distortion generator is a design that could
easily be accepted for a replacement to screens. This new design retains many of the
positive aspects of screens, including the adaptability of distortion patterns and the
possible precision to which patterns can be created. But, the split airfoil concept is more
efficient in that it requires less labor to conduct many different tests. It is also more
versatile in that, unlike screens, it can produce both steady state and transient distortion
effects. With more testing and development, the Split Airfoil Total Pressure Distortion
Generator could become a very powerful tool for gas turbine engine test and evaluation.
85
Appendix A: Force Plots for Mach 0.6 Flow
0
50
100
150
200
250
0 20 40 60 80 100
Angle in degrees
Dra
g Fo
rce
(lbf) 1 by 1 inch
2 by 2 inch3 by 3 inch4 by 4 inch5 by 5 inch
Figure A-1: Drag forces in Mach 0.6 flow for multiple sizes of square wedges.
0.00
50.00
100.00
150.00
200.00
250.00
300.00
0 20 40 60 80 100
Angle in Degrees
Lift
Forc
e (lb
f) 1 by 1 inch2 by 2 inch3 by 3 inch4 by 4 inch5 by 5 inch
Figure A-2: Lift forces in Mach 0.6 flow for multiples sizes of square wedges.
86
Appendix B: Force Plots for Mach 0.1 Flow
0
1
2
3
4
5
6
7
0 20 40 60 80 100Angle in Degrees
Pres
sure
For
ce (l
bf)
1 by 1 inch2 by 2 inch3 by 3 inch4 by 4 inch5 by 5 inch
Figure B-1: Pressure forces in Mach 0.1 flow for multiples sizes of square wedges.
01122334455
0 20 40 60 80 100
Angle in degrees
Dra
g Fo
rce
(lbf) 1 by 1 inch
2 by 2 inch3 by 3 inch4 by 4 inch5 by 5 inch
Figure B-2: Drag forces in Mach 0.1 flow for multiple sizes of square wedges.
87
0.00
1.00
2.00
3.00
4.00
5.00
6.00
0 20 40 60 80 100Angle in Degrees
Lift
Forc
e (lb
f) 1 by 1 inch2 by 2 inch3 by 3 inch4 by 4 inch5 by 5 inch
Figure B-3: Lift forces in Mach 0.1 flow for multiples sizes of square wedges.
88
Appendix C: Free-Body Diagrams for Wedge
F p
F act
F L
F D
x
y
Reaction Forces
1/2 Wedge
A θ
Figure C-1: Free-body diagram for front-supported wedge.
F act
F sup
F P
F D
y
x Reaction Forces
1/2 Wedge
A θ
f(θ)
Figure C-2: Free-body diagram for rear-supported wedge.
89
F act
F sup
F P F L
F D
x
y
Reaction Forces
1/2 Wedge
A θ
Figure C-3: Free-body diagram for center-supported wedge.
90
Appendix D: Free-Body Diagrams for Actuation
��������������������������������
����������������
�������������������������
Fp
FD
FL
Fp FL
FD
x
yQ
FlowFdir
Fdir
Figure D-1: Free-body diagram for rotationally actuated wedge.
������������������������������������
���������������
������������������������������
Fp
FD
FL
Fp FL
FD
x
yQFlow
Fdir
Figure D-2: Free-body diagram for horizontally actuated wedge.
91
�������������������������������������
��������������������
������������������������������
Fp
FD
FL
Fp FL
FD
Fdir
Fdir
x
yQ
Flow
Figure D-3: Free-body diagram for vertically actuated wedge.
Figure E-3: Force comparison for flow of Mach 0.1.
0
0.5
1
1.5
2
2.5
0 20 40 60 80 100 120 140 160 180 200
Angle (degrees)
Forc
e (lb
f)
dragliftpressureactuationviscous
Figure E-4: Force comparison for flow of Mach 0.08.
94
Appendix F: Uncertainty Analysis An uncertainty analysis was performed on the Mach number in order to assess the
error in the four repetitions of testing. The Mach number was determined from equation
F-1, which originates from the Mach number definition of the ratio of velocity to sound
speed.
TR
p
M**
*5.0γ
ρ∆
= (F-1)
where: ∆p = pressure change ρ = density of air γ = specific heat ratio R = gas constant of air T= temperature of air
The error of each component propagates through this equation. The error of the
manometer reading, ∆p, is ± 0.001 in H2O. The error of the density comes from the error
of the barometer reading the pressure, ± 0.0041 in H2O, and the thermometer, ± 0.05oC.
Based on the ideal gas law, this gives a density error of 0.0373 kg/s. Using equation F-2,
a Mach number of 0.100, ∆p of 3.000 in H2O, a density of 1.2300 kg/s, and a temperature
of 26.0oC, the overall Mach number error was found to be ± 0.015.
)(**
)(*5.0
error
error
error
error TTR
pp
MM+
+∆+∆
=+γ
ρρ (F-2)
This result shows that the first three separate runs were equivalent, within the error of the
tests. The fourth run was just outside the error of the tests, and actually acts as a separate
test condition. Therefore, non-dimensionalizing the results not only negated the effects
of error to the measurements but also allowed comparison between separate test
conditions.
95
References
Adamczyk, J.J., “Unsteady Fluid Dynamic Response of an Isolated Rotor with Distorted Inflow,” AIAA Paper No. 74-49, 1974. Alford, J.S., “Inlet Flow Distortion Index,” International Days of Aeronautical Sciences, ONERA, May 1957. Ashby Jr., G.C., “Investigation of the Effect of Velocity Diagram on Inlet Total-Pressure Distortion Through Single-Stage Subsonic Axial-Flow Compressors,” NACA RM L57A03, 1957. Beale, D., private correspondence, 1997. Beers, F.P., Johnston Jr., E.R., Mechanics of Materials, 2nd Ed., New York, McGraw-Hill, Inc.:1992. Blevins, Robert D., Applied Fluid Dynamics Handbook, New York, Van Norstrand Reinhold Company: 1984. Boller, S.M., “One-Dimensional Dynamic Wake Response in an Isolated Rotor Due to Inlet Total Pressure Distortion,” M.S. Thesis, Virginia Polytechnic Institute and State University (Blacksburg, Virginia, 1998). Braithwaite, W.M., Garber Jr., E.J., and Mehalic, C.M., “The Effect of Inlet Temperature and Pressure Distortion on Turbojet Performance,” AIAA Paper No. 73-1316, 1973. Brimlow, B., Collins, T.P., and Pfefferkorn, G.A., “Engine Testing in a Dynamic Environment,” AIAA Paper 74-1198, October, 1974. Campbell, Wilfred, “Protection of Turbine Disk Wheels from Axial Vibration,” Proceedings of the Cleveland Spring Meeting, ASME, May 1924. Carta, F.O., “Analysis of Unsteady Aerodynamic Effectos on Axial Flow Compressor Stage with Distorted Inflow.” Project Squid—Technical Rep. UARL—A—UP, 1972. Carta, F.O., “Aeroelasticity and Unsteady Aerodynamics,” Aircraft Propulsion Systems Technology and Design, Editor: Gordon C. Oates, (Washington D.C.: American Institute of Aeronautics and Astronautics, Inc., 1989.) Childs, J.H., Kochendorger, F.D., Lubick, R.J., and Friedman, R., “Stall and Flame-Out Resulting from Firing of Armament,” NACA RM E55E25, 1955.
96
Colpin, J., Kool, P., “Experimental Study of an Axial Compressor Rotor Transfer Function with Non-Uniform Inlet Flow,” ASME 78-GT-69, April 9-13, (London, England, 1978.) Cotter, H.N., “Integration of Inlet and Engine—an Engine Man’s Point of View,” SAE Paper 680286, April-May, 1968. Cousins, W.T., “Techniques for the Analysis of Unsteady Pressure Measurements in an Axial-Flow Compressor.” M.S. Thesis, Virginia Polytechnic Institute & State University, (Blacksburg, Virginia, 1979.) Cousins, W.T., O’Brien, W.F., “Axial-Flow Compressor Stage Post-Stall Analysis,” AIAA Paper No. 85-1349, 1985. Danforth, C.E., “Distortion-Induced Vibration in Fan and Compressor Blading,” Journal of Aircraft, vol. 12, no. 4, April 1975, pp. 216-225. Datko Jr., J.T., O’Hara, J.A., “The Aeromechanical Response of an Advanced Transonic Compressor to Inlet Distortion,” ASME 87-GT-189, 1987. Davis, M., Hale, A., Beale, D., “An argument for Enhancement of the Current Inlet Distortion Ground Test Practice for Aircraft Gas Turbine Engines,” ASME 2000-GT-0505, 2000. DiPietro, Tony, “Fundamental Wind Tunnel Experiments for Total Pressure distortion Generator Concept Selection,” Year End Report for Sverdrup Technology, January, 1996. Eddy, G., “Study of Steady-State Wake Characteristics of Variable Angle Wedges,” M.S. Thesis, Virginia Polytechnic Institute & State University, (Blacksburg, Virginia, 2001.) Fleeter, S., Jay, R.L., Bennett, W.A., “Rotor Wake Generated Unsteady Aerodynamic Response of a Compressor Stator,” ASME 78-GT-112, 1978. Greitzer, E.M., “Surge and Rotating Stall in Axial Flow Compressors; Part I: Theoretical Compression System Model.” ASME 75-GT-9, 1975. Hah, C., Rabe, D.C., Sulliavan, T.J., Wadia, A.R., “Effects of Inlet Distortion on the Flow Fieild in a Transonic Compressor Rotor,” ASME Journal of Turbomachinery, Vol. 120, April 1998, pp. 233-246. Hale, A.A., O’Brien, W.F., “A 3-D Turbine Engine Analysis Compressor Code (TEACC) for Steady-State Inlet Distortion.” Ph.D. Dissertation, Virginia Polytechnic Institute & State University, (Blacksburg, Virginia, 1996.) Henderson, R.E., Shen, I.C., “The Influence of Unsteady Rotor Response on a Distorted Flow Field.” ASME 81-GT-185, 1981.
97
Hurad, J., “Determination Experimentale des Lois de Transfert de Perturbations a la Traversee d’un Compresseur Axial,” AGARD CP-400, 68th Specialists Meeting, (Munich, Germany, September 8-9, 1986.) Jumael J., King P.S., O’Brien W.F, “Transient Total Pressure Distortion Generator Development” Final Report for Academic Qualification, (Blacksburg VA, July 1999.) Kimzey, W.F., “An Analysis of the Influence of Some External Disturbances on the Aerodynamic Stability of Turbine Engine Axial Flow Fans and Compressors.” AEDC-TR-77-80, August 1977. Ladd, J.A., Norby, W.P., “Dynamic Inlet Distortion Predictions Using CFD/Distortion Synthesis Approach,” AIAA Paper No. 98-2735, 1998. Lecht,M., Weyer, H.B., “On the Unsteady Aerodynamic Rotor Blade Loading in a Transonic Axial Flow Compressor with Unsteady Sate Inlet Distortion,” IUTAM Symposium on Aeroelasticity in Turbomachines, (Paris, October 18-23, 1976), pp. 1-16. Lotter, K.W., Jorg, J., “The Effect of Intake Flow Disturbances on APU Compressor Blade High Cycle Fatigue in the Airbus A300,” ICAS-82-4.6.2, (Seattle, Aug. 1982.) Luedke, J., “Use of Nonlinear Volterra Theory in Predicting the Propagation of Non-uniform Flow Through an Axial Compressor,” M.S. Thesis, Virginia Polytechnic Institute & State University (Blacksburg, Virginia, 2001.) Mabie, Hamilton H., Reinholtz, Charles F., Mechanisms and Dynamics of Machinery, Fourth Edition (New York: John Wiley & Sons, Inc, 1987.) Manwaring, S.R., Fleeter, S., “Inlet Distortion Generated Periodic Aerodynamic Rotor Response,” ASME 89-GT-299, 1989. Manwaring, S.R., Rabe, D.C., Lorence, C.B., Wadia, A.R., “Inlet Distortion Generated Forced Response of a Low Aspect-Ratio Transonic Fan,” ASME Journal of Turbomachinery, Vol. 119, October, 1996, pp. 665-676. Mazzawy, R.S., “Multiple Segment Parallel Compressor Model for Circumferential Flow Distortion.” ASME Journal of Engineering for Power, April, 1977. Mikolajxzak, A.A., Pfeffer, A.M., “Methods to Increase Engine Stability and Tolerance to Distortion,” AGARD LS No. 72, September 7, 1974. Mokelke, H., “Prediction Techniques,” AGARD LS No. 72, Sect. 5, 1974.
98
Munson, Bruce R., Young, Donald F., Okiishi, Theodore H., Fundamentals of Fluid Mechanics, Third Edition (New York: John Wiley & Sons, Inc, 1998.) Nagano, S., Takata, H., “Nonlinear Analysis of Rotating Stall,” Institute of Space and Aeronautical Science, University of Tokyo, Report No. 449, 1970. Newark Electronics Catalog 119, 2001-02. Overall, B.W., “Evaluation of an Airjet Distortion Generator Used to Produce Steady-State Total-Pressure Distortion at the Inlet of Turbine Engines,” AEDC-TR-76-141, December, 1976. Peacock, R.E., Overli, J., “Dynamic Flows in Compressors with Pressure Maldistributed Inlet Conditions,” AGARD CP-177, 1976. Pearson, H., McKenzie, A., “Wakes in Axial Compressors.” Journal of the Royal Aeronautical Society), p. 48, Vol. 63, July, 1959 Plourde, G.A., and Brimelow, B., “Pressure Fluctuations Cause Compressor Instability,” Proceedings of the Air Force Airframe-Propulsion Compatibility Symposium, Rept. AFAPL-TR-69-103, (Air Force Aero Propulsion Laboratory, Wright-Patterson AFB, OH, June 1970.) Rabe, D., Bolcs, A. and Russler, P., “Influence of Inlet Distortion on Transonic Compressor Blade Loading,: AIAA Paper No. 95-2461, July 1995. Rabe, D.C., Williams, C., Hah, C., “Inlet Flow Distortion and Unsteady Blade Response in a Transonic Axial-Compressor Rotor,” ASABE 99-7297, 1999. Reid, C., “The Response of Axial Flow Compressors to Intake Flow Distortion,” ASME 69-GT-29, 1969. Roberts, F., Plourde, G.A., Smakula, F. “Insights into Axial Compressor Response to Distortion.: AIAA Paper No. 68-565, 1968. SAE Aerospace Information Report -1419, Inlet Total-Pressure-Distortion Considerations for Gas-Turbine Engines, (Warrendale, PA, Rev. A, 1999). SAE Aerospace Recommended Practice -1420, Gas Turbine Engine Inlet Flow Distortion Guidelines, (Warrendale, PA, Rev. A, 1998). SAE Aerospace Resource Document -50015, A Current Assessment of the Inlet/Engine Temperature Distortion Problem, (Warrendale, PA, 1991.) Sexton, M.R., O’Brien, W.F., “A Model for Dynamic Loss Response in Axial-Flow Compressors,” ASME 81-GT-154, 1981.
99
Schwartz, J.R., “An Experimental and Analytical Investigation of Dynamic Flow Response of a Fan Rotor with Distorted Inlet Flow,” M.S. Thesis, Virginia Polytechnic Institute & State University, (Blacksburg, Virginia, 1999.) Slemon, Gordon R., Magnetoelectric Devices, John Wiley & Sons, Inc., (New York, 1966.) pp. 73-75. Small, M., “Improved Methods for Predicting the Effects of Inlet Flow Distortion on the Performance of Axial Flow Compressors,” M.S. Thesis, Virginia Polytechnic Institute and State University, (Blacksburg, Virginia, 2001.) Steenken, W.G., “Engine Operability,” Aircraft Propulsion Systems Technology and Design, Editor: Gordon C. Oates, (Washington D.C.: American Institute of Aeronautics and Astronautics, Inc., 1989.) Vuillez, C., Petot, B., “New Methods, New Methodology, Advanced CFD in the SNECMA Turbomachinery Design Process,” AGARD LS No. 195, May 1994. Walter, W.A. and Shaw, M., “Predicted F100 Engine Response to Circumferential Pressure and Temperature Distortion,” 1979 Wells, H.S., “A Study of Rocket Exhaust Gas Ingestion Simulation Techniques for the AEDC,” ARO ENGR RPT 77-3, March 1977. www.festo-usa.com/products/allstar.html www.physikinstrumente.com Younghans, J.L., Paul, D.L., “Inlets and Inlet/Engine Integration,” Aircraft Propulsion Systems Technology and Design, Editor: Gordon C. Oates, (Washington D.C.: American Institute of Aeronautics and Astronautics, Inc., 1989.)
100
Vita
Kevin B. Cramer
The author, son of Richard Jr. and Mary-Catherine Cramer, was born in 1978 in
Clearwater, Florida. As a child he moved up the east coast finding his way to Richmond,
Virginia. He attended high school at Midlothian High School. His decision to become
an engineer led to his enrollment in Mechanical Engineering at Virginia Tech in the fall
of 1998. At this time he also enrolled in Air Force Reserve Officers Training Corps.
Graduating from Virginia Tech in December of 2000, he was commissioned a Second
Lieutenant in the United States Air Force. In the spring of 2001, he began his graduate
studies in Mechanical Engineering, again at Virginia Tech. Upon graduation, he will